Data 2401 - Final Report
Tuan Pham
2023-04-26
Why is this subject?
A controversial topic (Gun Violence)
Provide an edge of the problems, a tip of the iceberg.
Raise your insights about the problems, and then you can draw conclusions for yourself in order to protect you and your family.
Recognize the primitive signs, not to live in skeptics but to enhance your awareness.
“If you look at the number of Americans killed since 9/11 by terrorism, it’s less than 100. If you look at the number been killed by gun violence, it’s in the tens of thousands.”
DATA
Words on my data set
164 observations and 26 variables from 1960 to 2023. (It was 143 observations and 25 variables originally from 1982 to 2022)
My data set’s limitation:
- Two data sets with two completely different structures.
- Various manually plug-in data (15%)
Mother Jones
https://www.motherjones.com/politics/2012/12/mass-shootings-mother-jones-full-data/
The data set’s authors are MARK FOLLMAN, GAVIN ARONSEN, and DEANNA PAN.
The time span they collect is from 1982 to present, which included
originally 143 observations and 25 variables. This is my work’s mainframe.The Violence Project
No specific author’s names.
The time span is from 1960 to 2022. This is the supplemental data for my main data frame.
DATA (cont.)
my_data <- read.csv('Mother Jones - Mass Shootings Database, 1982 - 2023 - Sheet1.csv', na.strings = "-")
my_data2 <- read.csv('Violence Project Mass Shooter Database - Version 6.1 - Full Database.csv', na.strings = "-")both in form of
csv file.First set: familiar with what we have been learned so far.
Second set: highly complicated
Missing data:
NA (Not Available) - Missing value
Unclear - The information has not been revealed by the authority.
TBD (To Be Determined) - The information has not been confirm yet.
A Glimpse over the Data
glimpse(my_data)## Rows: 164
## Columns: 26
## $ case <chr> "Louisville bank shooting", "Nashvill…
## $ city <chr> "Louisville", "Nashville", "East Lans…
## $ state <chr> "KY", "TN", "MI", "CA", "CA", "VA", "…
## $ date <chr> "4/10/2023", "3/27/2023", "2/13/2023"…
## $ summary <chr> "Connor Sturgeon, 25, opened fire ins…
## $ fatalities <int> 5, 6, 3, 7, 11, 6, 5, 3, 5, 3, 7, 3, …
## $ injured <int> 8, 6, 5, 1, 10, 6, 25, 2, 2, 2, 46, 0…
## $ total_victims <int> 13, 12, 8, 8, 21, 12, 30, 5, 7, 5, 53…
## $ location <chr> "workplace", "School", "School", "wor…
## $ age_of_shooter <int> 25, 28, 43, 67, 72, 31, 22, 22, 15, 2…
## $ prior_signs_mental_health_issues <chr> "Yes", NA, NA, NA, "Yes", NA, "Yes", …
## $ mental_health_details <chr> NA, NA, NA, NA, "According to the LA …
## $ weapons_obtained_legally <chr> "Yes", "Yes", "Yes", NA, NA, NA, NA, …
## $ where_obtained <chr> "gun dealership in Louisville", NA, N…
## $ weapon_type <chr> "Semiautomatic Rifle", "One Semiautom…
## $ weapon_details <chr> "AR-15 rifle", NA, NA, NA, NA, NA, NA…
## $ race <chr> "White", "White", "Black", "Asian", "…
## $ gender <chr> "M", "F (\"identifies as transgender\…
## $ sources <chr> "https://apnews.com/article/downtown-…
## $ mental_health_sources <chr> NA, NA, NA, NA, "https://www.latimes.…
## $ sources_additional_age <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ latitude <dbl> NA, NA, NA, NA, NA, 36.77262, 38.8809…
## $ longitude <dbl> NA, NA, NA, NA, NA, -76.25128, -104.7…
## $ type <chr> "Mass", "Mass", "Mass", "Spree", "Mas…
## $ year <int> 2023, 2023, 2023, 2023, 2023, 2022, 2…
## $ day_of_week <chr> "Monday", "Monday", "Monday", "Monday…Data Key Terms
Case: Case’s name well-known by the media
City/State: Location where the incidents happened
Date/Year/Date of Week: Specific day and year when the incidents occurred
Summary: Summary about the case
Fatalities/Injured/Total Victims: Facts that stand out from the incident
Location: Type of location where the incidents occurred
Age of Shooter/Race/Gender/Prior Sign Mental Issues: The shooter profile
Weapons Obtained Legally/Where Obtained/Weapons Type/Weapons Details: Weapons profile
Sources/Mental Health Sources/Additional Age Source: All sources that were used to conduct this data set.
Longitude/Latitude: GPS coordination of the incident’s location
Type: Mass Shooting or Shooting Spree designated to the incident
Packages
The tidyverse.
Data wrangling and plots.
#install.packages(tidyverse)
library(dplyr)
library(tidyverse)Data Wrangling
All the data type was character, even with numeric variables’ types
There is a limitation in term of data information about the observations of
locationvariable in the raw data. Most of the Nightlife observations like Bar/Club/Restaurants was classified asOther, which make the original data having a large quantity of observations inOthercategory.
# Assign the variables to the data type of my choice.
my_data$age_of_shooter <- as.integer(my_data$age_of_shooter)
my_data$fatalities <- as.integer(my_data$fatalities)
my_data$injured <- as.integer(my_data$injured)
my_data$total_victims <- as.integer(my_data$total_victims)
my_data$latitude <- as.numeric(my_data$latitude)
my_data$longitude <- as.numeric(my_data$longitude)# remove newline
my_data$location <- str_replace_all(my_data$location,'[\r\n]','')
my_data$race <- str_replace_all(my_data$race,'[\r\n]','')
# replace a string by another one
my_data$location[my_data$location == 'religious' | my_data$location == 'Religious'] <- 'Religious Place'
my_data$location[my_data$location == 'workplace'] <- 'Workplace'
my_data$race[my_data$race == 'unclear'] <- 'Unclear'
my_data$race[my_data$race == 'black'] <- 'Black'
my_data$race[my_data$race == 'white'] <- 'White'
my_data$gender[2] <- 'Trans'
my_data$gender[my_data$gender == 'M'] <- 'Male'
my_data$gender[my_data$gender == 'F'] <- 'Female'What’s The Mass Shooting?
The FBI defines
a mass shootingas any incidents in which at least four people are murdered with a gun.
What questions should be raised?
Will the time frame would say anything about the incidents
in general?Would the
age, race, and gendergive any insights about the shooter’s profile?What would stand out if we cross
the shooter with prior mental health issuesout of the equation?Where are the
locationsthat the incidents likely take place?What
types of weaponsthe assaiants likely use?What conclusion about the
age of shooter,raceandprior mental health issuescould we draw?What is interesting about the connection between
age of the shooteroveryear?Will
genderplay any roles in corresponding toage of the shooter?How have the incidents distributed
across the America?
A First Glance about the Incidents over Years
Data Summary
## age_of_shooter fatalities injured total_victims
## Min. :11.0 Min. : 3.0 Min. : 0.00 Min. : 3.00
## 1st Qu.:23.0 1st Qu.: 4.0 1st Qu.: 1.00 1st Qu.: 6.00
## Median :32.0 Median : 6.0 Median : 3.00 Median : 10.00
## Mean :33.9 Mean : 7.5 Mean : 10.56 Mean : 18.09
## 3rd Qu.:43.0 3rd Qu.: 8.0 3rd Qu.: 9.50 3rd Qu.: 16.00
## Max. :72.0 Max. :58.0 Max. :546.00 Max. :604.00
## NA's :1 NA's :11. Las Vegas Strip Massacre: 604 victims
2. LA Dance Studio Mass Shooting: Oldest age for a mass shooter
3. West Middle School Killings: Youngest age
- Those three cases are the outliers for each of the variables that they represent.
Who are they?
The Picture in General
Before 2002
After 2002
Since 2002 is the year without any major incidents about the mass shooting, I chose it as a reference point for my split stats.
Before 2002, the story seems to be about some certain races, but after 2002, it becomes all the races’ story.
Remember there are four decades before 2002, and only two decades after 2002, but the cases after 2002 shoot up more than double before 2002, 55 versus 109 respectively.
The Average Age of the Shooters among Races
| Race | Average Age of the Shooters |
|---|---|
| White | ~ 28-29 years old |
| Latino | ~ 32-33 years old |
| Black | ~ 38-39 years old |
| Asian | ~ 41 years old |
| Native Am. | ~ 18 years old |
Gender of the Shooters
| Gender | Percentage |
|---|---|
| Male | 97% |
| Female | 2.5% |
| Transgender | 0.5% |
Gender and Age of the Shooters
| Gender | Most Likely |
|---|---|
| Male | Early 20’s to Mid 40’s |
| Female | Mid 20’s or Mid 40’s |
Where Do The Mass Shootings Likely Occur?
| Location | Frequency |
|---|---|
| Workplace | ~ 34% |
| School | ~ 16% |
| Bar/Club/Rest. | ~ 11% |
| Retail | ~ 10% |
| Other | ~ 9% |
| Religious Place | ~ 6% |
It is heartbroken to see School is second rank on the list, which means a lot of innocent kids got their future ahead taken.
What Weapons Were Likely Used by The Assailants?
| Firearm | Percent of Carrying |
|---|---|
| Semi-Auto Handgun | ~ 41% |
| Semi-Auto Rifle | ~ 20% |
| Handgun(Old Versions) | ~ 6.7% |
| Rifle(Old Version) | ~ 5% |
| Assault Rifle | ~ 5% |
| Shotgun | ~ 4% |
- The preferred weapons was used by the mass shooter are Semi-Auto Handgun, Semi-Auto Rifle or both of them.
- They are light weight, high bullet rate and big magazines. It could help the shooters cause mass casualties in a short period of time.
Race and Mental Health Issues
| Race | Prior Mental Health Issues |
|---|---|
| Asian | 90% |
| White | 69% |
| Latino | 67% |
| Black | 43% |
Will The Shooter’s Age Be Related over Time?
At the first glance, we are not able to say anything about the relationships between the ages of the shooters over the time. Hence I decide to divide the graph into the zone and sub-areas like the graph below.
+There are 3 things we can spot instinctively that:
The right-hand side of the first zone and the third zone yield more dots than the whole bigger left-hand side one. Those zones are the least expected ages for the incidents to happen, but they still have more cases than its left-hand side ones.
The density of the right-hand side in the second zone is more condense than the left-hand side, even the area is smaller but the dots represent heavier than the other side. Zone II is the main labor force of the market, and it will be divided into three sub-areas.
Take a look back to the whole picture, the right-hand side seems to be outnumbering to the left hand side, which means the number of cases has rapidly increased for last two decades.
Take a closer look to the Zone II; I divided it into three sub-zones:
Sub-area 1: Young people with no experiences either in life or at work tend to get shocked and get impulses if they got history of abuses. They got full energy to do what they want to do. They will make tremendous mistakes with that source of energy unless they have well-guidance from older adults.
Sub-area 2: Most of people in this area got a career to follow, a family to take care of and some life’s experiences. Health is at peak. Some of the conflicts at work might trigger their impulses, but in general they can hold it through.
Sub-area 3: In this age’s range, an adult is mature in life both financially and mentally, even the health is in downturn. Kids got older, career is more stable. Anybody in this age got something to lose, so they normally do not do anything irrationally unless they got some mental health issues. That probably explains why there are no shooters without prior mental health in this range.
Zone II is the area where the incident occurs the most. Since this age’s range is the main force in the labor market, which explains why the mass shooting cases at the workplace appear to be the most.
With Prior Mental Health Issues
Now, the shooters with the signs of the prior mental issues will be added to the graph.
The incidents in which the shooter had prior mental health issues have plotted as the plus sign (+) on the plot above.
Without Prior Mental Health Issues Plot
Now we take those cases out of the plot to see how the current plot looks like.
Compare to the original plot, we can see intuitively the dots’ density was reduced significantly. Hence, we are going to find the difference between with and without prior mental health issues by numbers.
Realize Intuitive Consideration by Numbers
We can draw some remarks by spotting the plots. Now we are going to
consider some numbers from the data to see how much the
mental health issues contribute to the problem.
my_data %>%
filter(prior_signs_mental_health_issues == "Yes") %>%
group_by(year) %>%
nrow()## [1] 80If we filter out the cases with the prior mental health issues, there are eighty cases was off the chart, which is almost half of cases of mass shooting in the US since 1960.
my_data %>%
filter(weapons_obtained_legally == "No") %>%
group_by(year) %>%
nrow()## [1] 16In a different case, I cross off the
legal weapons obtained, only 16 cases was off the chart,
which roughly 10% of all of the cases.
Conclusion:
People always argued about either we should do the background check or adjust the law over gun control. Now we can state that background check is more important than gun control, especially background check on mental health issues is crucial. Decreasing the cases down to fifty percent is ideal, but twenty or thirty percent down is sufficient to save many lives.
| Prior Signs of Mental Health Issues | Age of the Shooter |
|---|---|
| Yes | around 23 and 40 |
| No | around 30 |
Geography Graph of The Events
The graph shows us an idea that the incidents most likely occurs over the East and West side of the country, and the Mid-west is least likely to happen the mass shootings.
Colorado is the state in top 5 rating of mass shooting even the population rank is not in top 20 nationwide.
Massachusetts surprisingly has no records on mass shooting even the population is in top 16 nationwide.
State Population Source: https://www.statsamerica.org/sip/rank_list.aspx?rank_label=pop1
Pick Your Day to Go Out.
Sunday is the deadliest day of the week in term of Mass Shooting but Monday is the most likely day for the Mass Shooter plan to act.